Who Makes Trends? Understanding Demographic Biases in Crowdsourced Recommendations

By Abhijnan Chakraborty, Johnnatan Messias, Fabrício Benevenuto, Saptarshi Ghosh, Niloy Ganguly, Krishna P. Gummadi. 2017

78 citations recorded by the source corpus.

Users of social media sites like Facebook and Twitter rely on crowdsourced content recommendation systems (for example, Trending Topics) to retrieve important and useful information. Contents selected for recommendation indirectly give the initial users who promoted (by liking or posting) the content an opportunity to propagate their messages to a wider audience. Hence, it is important to understand the demographics of people who make a content worthy of recommendation, and explore whether they are representative of the media site's overall population. In this work, using extensive data collected from Twitter, we make the first attempt to quantify and explore the demographic biases in the crowdsourced recommendations. Our analysis, focusing on the selection of trending topics, finds that a large fraction of trends are promoted by crowds whose demographics are significantly different from the overall Twitter population. More worryingly, we find that certain demographic groups are systematically under-represented among the promoters of the trending topics. To make the demographic biases in Twitter trends more transparent, we developed and deployed a Web-based service Who-Makes-Trends at twitter-app.mpi-sws.org/who-makes-trends.

Subfields: Communication, Sociology and Political Science, Statistical and Nonlinear Physics, Recommender Systems, AI Ethics, Fairness & Society

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Communication

1. Alexandra Olteanu
2. Fernando Diaz
3. Gordon Pennycook
4. David G. Rand
5. Carlos Castillo
6. Emre Kıcıman
7. Jungseock Joo
8. Kimmo Kärkkäinen

In Sociology and Political Science

1. Alexandra Olteanu
2. David G. Rand
3. Gordon Pennycook
4. Fernando Diaz
5. Carlos Castillo
6. Emre Kıcıman
7. Kimmo Kärkkäinen
8. Jungseock Joo

In Statistical and Nonlinear Physics

1. Alexandra Olteanu
2. Fernando Diaz
3. Carlos Castillo
4. Emre Kıcıman
5. David G. Rand
6. Gordon Pennycook
7. Fabrício Benevenuto
8. Jungseock Joo

In Recommender Systems

1. Fernando Diaz

In AI Ethics, Fairness & Society

1. Alexandra Olteanu
2. Fernando Diaz
3. Carlos Castillo
4. David G. Rand
5. Gordon Pennycook
6. Emre Kıcıman
7. Angelina Wang
8. Olga Russakovsky

Citations in the corpus

• FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age (2019, background): “Notable examples are demographic analyses of social media users using their photographs [9, 43, 64, 65, 62].”
• Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries (2019, background): “…they are central to a research study, in others their use is more subtle or peripheral: e.g., a study analyzing public opinion or popularity across demographics may infer those demographics using pre-trained models for facial recognition (e.g., Wang Y. et al., 2016b; Chakraborty et al., 2017).”
• FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age for Bias Measurement and Mitigation (2021, background): “Notable examples are demographic analyses of social media users using their photographs [9, 45, 65, 66, 63].”
• Demographics of News Sharing in the U.S. Twittersphere (2017, background): “Besides that, the researchers show that the strategy of getting demographic data from Face++ is reliable and provides accurate demographic information for gender and race, encouraging the application of this strategy in other recent efforts [13].”
• Shifting attention to accuracy can reduce misinformation online (2021, methodology): “, 2016)), and age and gender using based on profile pictures using the face plus plus algorithm (An & Weber, 2016; Chakraborty et al., 2017; Kteily, Rocklage, McClanahan, & Ho, 2019).”